New Open Hardware ECU Compatible with rusefi

[openlunchbox]

We are going to start making an open source hardware ECU compatible with rusefi.

Details are here:
http://openlunchbox.com/open-ecu/

http://openlunchbox.com/smf/index.php?topic=82.0

The ECU IO modules will be able to use an STM32F4 or the new STMF7 microcontrollers to keep compatibility with the project here.

The micros are on a small carrier similar to the ST Discovery modules, only we are making the microcontroller modules and IO boards very ruggedized to handle the EMC, ESD and wide temperature range that automotive requires. There will also be carrier boards for other micros to suit other open source EFI projects. This way any of the open EFI software projects can take advantage of a common well made, low cost, open hardware platform.

[kb1gtt]

Do you have EMC testing capabilities? It sounds like you have detailed rusEFI's specs on this site. We have aimed to uphold the specs detailed there but have only been able to do some basic testing to validate the design.

I see several very ambitious project there, but I couldn't find much for a tangible result.

What do you plan for a case and harness connector? We have found that these connector and enclosures are very expensive. Often they cost more than the PCB and electronics that go inside the enclosure.

[openlunchbox]

We have just about everything. I've been making hardware for over 30 years and spend half my time in China sourcing and manufacturing.

The ambitious projects are not even listed on the site yet. I consider them one banana out of five banana projects (on a scale of 1 to 5 bananas, 5 being the most difficult)

Most OEM enclosures I've seen are extrusions with formed covers. These are pretty easy to make but nobody has asked for OEM style cases or connectors. What we have been asked for is a small unit for the 1-2cyl applications that is easily fit into the space available in a motorcycle. That why we have 3 different size IO boards that all use the same size and pinout microcontroller modules. One IO board and case for 1-2cyl applications, one for 1-6cyl and yet another for more cylinders.

Now is the time to chime in on the cases desired and the types of connectors that are wanted.

If you wish to chat in real time we have a channel on freenode IRC #openlunchbox

[AndreyB]

I've been making hardware for over 30 years and spend half my time in China sourcing and manufacturing.

Have you seen https://www.facebook.com/rusEfiECU/photos/a.264355967021383.1073741827.264354717021508/293632644093715/?

[abecedarian]

Well, this was a little unexpected turn, no? I mean "unexpected" as in it seems a little early to have your things co-opted out to some other project.

Andrey, I hope you receive product, without charge, so you can validate your code works against these boards, otherwise someone else owns it.

[openlunchbox]

I've been making hardware for over 30 years and spend half my time in China sourcing and manufacturing.

Have you seen https://www.facebook.com/rusEfiECU/photos/a.264355967021383.1073741827.264354717021508/293632644093715/?

No I haven't. Thanks for the link.

What I'm doing is not making a Frankenso board that uses a STM32F4 dev board but a complete ruggedized ECU that uses the STM32F4/7 for compatibility with rusefi or other microcontroller modules for compatibility with other open EFI projects.

What I require is an automotive grade ECU for aftermarket performance applications, and not a hobby grade stack of dev boards. People have real applications for an ECU that they can fully customize using open source software.

[kb1gtt]

When you have hardware that's significantly designed, I can offer to give it a review to help make it stronger.

What PCB layout tools are you planning to use? Could I encourage KICAD? Also have you seen jpcbsim which uses OpenEMS

http://openems.de/start/index.php

These tools can help make the design stronger.

Have you seen the general design specs that we are aiming to uphold found here http://rusefi.com/wiki/index.php?title=Manual:Hardware

Do you see weaknesses in the Frankenso board that does not meet you desired needs? I believe russian was inquiring if you might be interested in building the next spin of Frankenso. We aim to build that some time in the near-ish future. We have been holding out as I get a test mule to allow additional testing of the R0.2 board. It would be nice if we can find ways to combine efforts, instead of several independent efforts.

Have you seen the "brain board" forum thread found here http://rusefi.com/forum/viewtopic.php?f=4&t=381 also some day when we need more IO we have a potential plan to make a different IO board using this http://rusefi.com/forum/viewtopic.php?f=4&t=749

Other brain boards are possible. As well we plan for boards to be installed above the several Wxx jumpers near the harness connector. We can add features and change the pin out by adding a PCB above those Wxx jumpers.

We are using the discovery dev board for prototype prototype purposes. The long term plan is to either populate it on the Frankenso board natively, or to allow for different brain boards via headers as we are currently using with the discovery board.

About the general design requests,
-- I would suggest, an option to include a water tight connector that includes a harness with a pigtail.
-- I would suggest, an option to include a modular connector design that allow the potential for being connected to many different existing connectors and existing pinouts.
-- I would suggest, allowing external expansion options. On Frankenso we have a header that includes pins for 5V, CAN and GND. This allows for external board(s) to be added at a future date, potentially for items like accelerometers, wheel speed sensors, ect.
-- I would suggest, a switch mode power supply with an excess of current capability

Again if you get something that has been significantly developed, I can offer a review to help make it stronger.

[openlunchbox]

Thank you for the input. The time to chime in with feature requests is now during design vs after. It's not a very complex hardware design. No more than a few days here and there after locking in the features.

It's looks like we already have all your concerns covered.

Most often I use NX for parametric modelling, Flotherm and Radtherm for thermal analysis, Ansys for simulation and Cadence/Orcad Capture and Allegro for pcb layout.

[kb1gtt]

What Ansys do you have? HFSS, or other?

What version of OrCad do you use?

Can I encourage you to use some less pricey tools for schematic capture and PCB layout? It can help get others involved in an open source project. I have access to HFSS, Inventor, Autodesk CFD, ADS, and an older version of OrCAD. I use open and low cost tools when I work with people in open source projects as they'll then have the ability to make modifications and participate. I understand wanting to use the tools you already have access to and know how to work with.

[openlunchbox]

We run just about everything (Catia, Creo, SW, Altium, PADS, etc) except for some of the open source CAD tools. They just take far too long to get things done.

I typically post IGS, STP as well along with pdf, gerbers so people can import or convert them if you want to use Kicad, Eagle, BRL, FreeCAD etc.

Post your wish list for features here: http://openlunchbox.com/smf/index.php?topic=72.0

[kb1gtt]

I do not want to sign in to yet another forum and keep track of yet another password when I see little content. Once there is more significant content I may consider signing up, but I'm not interested until then.

Can you link me to a typically released project? I've only see empty shells of projects in github.

The only tool (open or commercial) that I know of that can bring in a gerber to a layout is KICAD. I have seen a 3rd party tool that can convert a gerber to a native layout, but I forget that webpage URL and I forget how much it costs to convert.

[openlunchbox]

Since the ECU can be compatible with rusefi or other EFI projects such as Libreems or FreeEMS by just swapping the microcontroller module, I'm keeping the design details on a neutral site. http://openlunchbox.com/smf/index.php#c7 and http://openlunchbox.com/open-ecu/

[kb1gtt]

I have no problems with you keeping the specs on pages you can control. I've done the same after getting burnt with o5e. Currently russian pulls from my repo to get his repo. I think you have all the major design specs covered. When you have something drafted, feel free to request my review. I'd be happy to give you some feedback and thoughts.

[kb1gtt]

About the open phone effort, or potentially other applications and modular design... which uses greybus at the core https://github.com/projectara/greybus

[puff]

can't find the proper thread for this.
http://hackaday.com/2017/01/05/hands-on-with-the-first-open-source-microcontroller/
or that's not enough for our needs?

[kb1gtt]

Hmm, how much heat does it make and what is it's maximum temperature? I guess what is it's lowest temperature as well? To bad it was developed with the paid for Altium, this will limit peoples abilities to inspect the chips internals. It would have been cool if they could have done it with something like Magic
http://opencircuitdesign.com/magic/

I wonder how long until someone has ported it to Magic.